@article{finzel2026_116862,
  title = {Toward trustworthy AI with integrative explainable AI frameworks},
  author = {Finzel, Bettina},
  year = {2026},
  pages = {20-45},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-116862x},
  url = {https://fis.uni-bamberg.de/handle/uniba/116862},
  language = {eng},
  abstract = {As artificial intelligence (AI) increasingly permeates high-stakes domains such as healthcare, transportation, and law enforcement, ensuring its trustworthiness has become a critical challenge. This article proposes an integrative Explainable AI (XAI) framework to address the challenges of interpretability, explainability, interactivity, and robustness. By combining XAI methods, incorporating human-AI interaction and using suitable evaluation techniques, the implementation of this framework serves as a holistic XAI approach. The article discusses the framework’s contribution to trustworthy AI and gives an outlook on open challenges related to interdisciplinary collaboration, AI generalization and AI evaluation.},
  keywords = {trustworthy AI, explainable AI, EU AI act, integrative XAI frameworks, XAI in medicine}
}

@article{holzinger2026_114970,
  title = {Toward human-level concept learning : Pattern benchmarking for AI algorithms},
  author = {Holzinger, Andreas and Saranti, Anna and Angerschmid, Alessa and Finzel, Bettina and Schmid, Ute and Mueller, Heimo},
  year = {2026},
  pages = {1-21},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-114970x},
  url = {https://fis.uni-bamberg.de/handle/uniba/114970},
  language = {eng},
  abstract = {Artificial intelligence (AI) today is very successful at standard pattern-recognition tasks due to the availability of large amounts of data and advances in statistical data-driven machine learning. However, there is still a large gap between AI pattern recognition and human-level concept learning. Humans can learn amazingly well even under uncertainty from just a few examples and are capable of generalizing these concepts to solve new conceptual problems. The growing interest in explainable machine intelligence requires experimental environments and diagnostic/benchmark datasets to analyze existing approaches and drive progress in pattern analysis and machine intelligence. In this paper, we provide an overview of current AI solutions for benchmarking concept learning, reasoning, and generalization; discuss the state-of-the-art of existing diagnostic/ benchmark datasets (such as CLEVR, CLEVRER, CLOSURE, CURI, Bongard-LOGO, V-PROM, RAVEN, Kandinsky Patterns, CLEVR-Humans, CLEVRER-Humans, and their extension containing human language); and provide an outlook of some future research directions in this exciting research domain.}
}

@inproceedings{knoblach2026_114752,
  title = {A Framework for Context-aware XAI in Human-AI Collaboration},
  author = {Knoblach, Judith and Finzel, Bettina and Schmid, Ute},
  year = {2026},
  booktitle = {Contextualizing Explanations : Proceedings of the 3rd TRR 318 Conference},
  pages = {24},
  publisher = {Bielefeld University Press},
  address = {Bielefeld},
  doi = {10.64136/deef2831},
  url = {https://fis.uni-bamberg.de/handle/uniba/114752},
  language = {eng}
}

@article{finzel2025_116722,
  title = {Toward trustworthy AI with integrative explainable AI frameworks},
  author = {Finzel, Bettina},
  year = {2025},
  journal = {Information technology : it},
  volume = {67},
  number = {1},
  pages = {20-45},
  publisher = {De Gruyter},
  address = {Berlin},
  issn = {1611-2776},
  doi = {10.1515/itit-2025-0007},
  url = {https://fis.uni-bamberg.de/handle/uniba/116722},
  language = {eng},
  abstract = {As artificial intelligence (AI) increasingly permeates high-stakes domains such as healthcare, transportation, and law enforcement, ensuring its trustworthiness has become a critical challenge. This article proposes an integrative Explainable AI (XAI) framework to address the challenges of interpretability, explainability, interactivity, and robustness. By combining XAI methods, incorporating human-AI interaction and using suitable evaluation techniques, the implementation of this framework serves as a holistic XAI approach. The article discusses the framework’s contribution to trustworthy AI and gives an outlook on open challenges related to interdisciplinary collaboration, AI generalization and AI evaluation.},
  keywords = {trustworthy AI, explainable AI, EU AI act, integrative XAI frameworks, XAI in medicine}
}

@inproceedings{kaidashova2025_113752,
  title = {Toward Simple and Robust Contrastive Explanations for Image Classification by Leveraging Instance Similarity and Concept Relevance},
  author = {Kaidashova, Yuliia and Finzel, Bettina and Schmid, Ute Maria},
  editor = {Springer Nature Switzerland},
  year = {2025},
  booktitle = {KI 2025: advances in artificial intelligence : 48th German Conference on AI, Potsdam, Germany, September 16-19, 2025, proceedings},
  pages = {32-46},
  publisher = {Springer Nature Switzerland},
  address = {Cham},
  isbn = {978-3-032-02812-9},
  doi = {10.1007/978-3-032-02813-6\_3},
  url = {https://fis.uni-bamberg.de/handle/uniba/113752},
  language = {eng}
}

@article{finzel2025_107762,
  title = {Current methods in explainable artificial intelligence and future prospects for integrative physiology},
  author = {Finzel, Bettina},
  year = {2025},
  pages = {513-529},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-1077626},
  url = {https://fis.uni-bamberg.de/handle/uniba/107762},
  language = {eng},
  abstract = {Explainable artificial intelligence (XAI) is gaining importance in physiological research, where artificial intelligence is now used as an analytical and predictive tool for many medical research questions. The primary goal of XAI is to make AI models understandable for human decision-makers. This can be achieved in particular through providing inherently interpretable AI methods or by making opaque models and their outputs transparent using post hoc explanations. This review introduces XAI core topics and provides a selective overview of current XAI methods in physiology. It further illustrates solved and discusses open challenges in XAI research using existing practical examples from the medical field. The article gives an outlook on two possible future prospects: (1) using XAI methods to provide trustworthy AI for integrative physiological research and (2) integrating physiological expertise about human explanation into XAI method development for useful and beneficial human-AI partnerships.},
  keywords = {Explainable Artificial Intelligence (XAI), Physiology, Explainability, Interpretability, Survey}
}

@article{finzel2025_107758,
  title = {Current methods in explainable artificial intelligence and future prospects for integrative physiology},
  author = {Finzel, Bettina},
  year = {2025},
  journal = {Pflügers Archiv : European journal of physiology},
  volume = {477},
  number = {4},
  pages = {513-529},
  publisher = {Springer},
  address = {Berlin ; Heidelberg},
  issn = {1432-2013},
  doi = {10.1007/s00424-025-03067-7},
  url = {https://fis.uni-bamberg.de/handle/uniba/107758},
  language = {eng},
  abstract = {Explainable artificial intelligence (XAI) is gaining importance in physiological research, where artificial intelligence is now used as an analytical and predictive tool for many medical research questions. The primary goal of XAI is to make AI models understandable for human decision-makers. This can be achieved in particular through providing inherently interpretable AI methods or by making opaque models and their outputs transparent using post hoc explanations. This review introduces XAI core topics and provides a selective overview of current XAI methods in physiology. It further illustrates solved and discusses open challenges in XAI research using existing practical examples from the medical field. The article gives an outlook on two possible future prospects: (1) using XAI methods to provide trustworthy AI for integrative physiological research and (2) integrating physiological expertise about human explanation into XAI method development for useful and beneficial human-AI partnerships.},
  keywords = {Explainable Artificial Intelligence (XAI), Physiology, Explainability, Interpretability, Survey}
}

@article{finzel2025_106099,
  title = {Human-Centered Explanations : Lessons Learned from Image Classification for Medical and Clinical Decision Making},
  author = {Finzel, Bettina},
  year = {2025},
  pages = {157-167},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-1060994},
  url = {https://fis.uni-bamberg.de/handle/uniba/106099},
  language = {eng},
  abstract = {To date, there is no universal explanatory method for making decisions of an AI-based system transparent to human decision makers. This is because, depending on the application domain, data modality, and classification model, the requirements for the expressiveness of explanations vary. Explainees, whether experts or novices (e.g., in medical and clinical diagnosis) or developers, have different information needs. To address the explanation gap, we motivate human-centered explanations and demonstrate the need for combined and expressive approaches based on two image classification use cases: digital pathology and clinical pain detection using facial expressions. Various explanatory approaches that have emerged or been applied in the three-year research project “Transparent Medical Expert Companion” are shortly reviewed and categorized in expressiveness according to their modality and scope. Their suitability for different contexts of explanation is assessed with regard to the explainees’ need for information. The article highlights open challenges and suggests future directions for integrative explanation frameworks.},
  keywords = {Explainable Artificial Intelligence, Digital healthcare, Human-centered explanations, Explanation expressiveness, Explanation frameworks}
}

@inproceedings{finzel2024_106515,
  title = {Near Hit and Near Miss Example Explanations for Model Revision in Binary Image Classification},
  author = {Finzel, Bettina and Knoblach, Judith and Thaler, Anna Magdalena and Schmid, Ute},
  year = {2024},
  booktitle = {Intelligent Data Engineering and Automated Learning – IDEAL 2024 : 25th International Conference, Valencia, Spain, November 20–22, 2024, Proceedings, Part II},
  pages = {260-271},
  publisher = {Springer Nature Switzerland},
  address = {Cham},
  isbn = {9783031777370},
  doi = {10.1007/978-3-031-77738-7\_22},
  url = {https://fis.uni-bamberg.de/handle/uniba/106515},
  language = {eng}
}

@misc{finzel2024_106516,
  title = {When a Relation Tells More Than a Concept : Exploring and Evaluating Classifier Decisions with CoReX},
  author = {Finzel, Bettina and Hilme, Patrick and Rabold, Johannes and Schmid, Ute},
  year = {2024},
  pages = {1-56},
  publisher = {arXiv},
  doi = {10.48550/arxiv.2405.01661},
  url = {https://fis.uni-bamberg.de/handle/uniba/106516},
  language = {eng}
}

@inproceedings{finzel2024_106518,
  title = {Explaining with Attribute-Based and Relational Near Misses : An Interpretable Approach to Distinguishing Facial Expressions of Pain and Disgust},
  author = {Finzel, Bettina and Kuhn, Simon P. and Tafler, David E. and Schmid, Ute Maria},
  year = {2024},
  booktitle = {Inductive Logic Programming : 31st International Conference, ILP 2022, Windsor Great Park, UK, September 28–30, 2022, Proceedings},
  pages = {40-51},
  publisher = {Springer Nature Switzerland},
  address = {Cham},
  isbn = {9783031556296},
  doi = {10.1007/978-3-031-55630-2\_4},
  url = {https://fis.uni-bamberg.de/handle/uniba/106518},
  language = {eng}
}

@article{finzel2024_106097,
  title = {Human-Centered Explanations : Lessons Learned from Image Classification for Medical and Clinical Decision Making},
  author = {Finzel, Bettina},
  year = {2024},
  journal = {Künstliche Intelligenz : KI ; Forschung, Entwicklung, Erfahrungen},
  volume = {38},
  number = {3},
  pages = {157-167},
  publisher = {Springer},
  address = {Berlin},
  issn = {1610-1987},
  doi = {10.1007/s13218-024-00835-y},
  url = {https://fis.uni-bamberg.de/handle/uniba/106097},
  language = {eng},
  abstract = {To date, there is no universal explanatory method for making decisions of an AI-based system transparent to human decision makers. This is because, depending on the application domain, data modality, and classification model, the requirements for the expressiveness of explanations vary. Explainees, whether experts or novices (e.g., in medical and clinical diagnosis) or developers, have different information needs. To address the explanation gap, we motivate human-centered explanations and demonstrate the need for combined and expressive approaches based on two image classification use cases: digital pathology and clinical pain detection using facial expressions. Various explanatory approaches that have emerged or been applied in the three-year research project “Transparent Medical Expert Companion” are shortly reviewed and categorized in expressiveness according to their modality and scope. Their suitability for different contexts of explanation is assessed with regard to the explainees’ need for information. The article highlights open challenges and suggests future directions for integrative explanation frameworks.},
  keywords = {Explainable Artificial Intelligence, Digital healthcare, Human-centered explanations, Explanation expressiveness, Explanation frameworks}
}

@article{schwalbe2024_98420,
  title = {A comprehensive taxonomy for explainable artificial intelligence : a systematic survey of surveys on methods and concepts},
  author = {Schwalbe, Gesina and Finzel, Bettina},
  year = {2024},
  pages = {3043-3101},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-984203},
  url = {https://fis.uni-bamberg.de/handle/uniba/98420},
  language = {eng},
  abstract = {In the meantime, a wide variety of terminologies, motivations, approaches, and evaluation criteria have been developed within the research field of explainable artificial intelligence (XAI). With the amount of XAI methods vastly growing, a taxonomy of methods is needed by researchers as well as practitioners: To grasp the breadth of the topic, compare methods, and to select the right XAI method based on traits required by a specific use-case context. Many taxonomies for XAI methods of varying level of detail and depth can be found in the literature. While they often have a different focus, they also exhibit many points of overlap. This paper unifies these efforts and provides a complete taxonomy of XAI methods with respect to notions present in the current state of research. In a structured literature analysis and meta-study, we identified and reviewed more than 50 of the most cited and current surveys on XAI methods, metrics, and method traits. After summarizing them in a survey of surveys, we merge terminologies and concepts of the articles into a unified structured taxonomy. Single concepts therein are illustrated by more than 50 diverse selected example methods in total, which we categorize accordingly. The taxonomy may serve both beginners, researchers, and practitioners as a reference and wide-ranging overview of XAI method traits and aspects. Hence, it provides foundations for targeted, use-case-oriented, and context-sensitive future research.},
  keywords = {Explainable artificial intelligence, Interpretability, Taxonomy, Meta-analysis, Survey-of-surveys, Review}
}

@inproceedings{finzel2024_95294,
  title = {Domain-Specific Evaluation of Visual Explanations for Application-Grounded Facial Expression Recognition},
  author = {Finzel, Bettina and Rieger, Ines and Kuhn, Simon and Schmid, Ute},
  year = {2024},
  pages = {31-44},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-952943},
  url = {https://fis.uni-bamberg.de/handle/uniba/95294},
  language = {eng},
  abstract = {Research in the field of explainable artificial intelligence has produced a vast amount of visual explanation methods for deep learning-based image classification in various domains of application. However, there is still a lack of domain-specific evaluation methods to assess an explanation’s quality and a classifier’s performance with respect to domain-specific requirements. In particular, evaluation methods could benefit from integrating human expertise into quality criteria and metrics. Such domain-specific evaluation methods can help to assess the robustness of deep learning models more precisely. In this paper, we present an approach for domain-specific evaluation of visual explanation methods in order to enhance the transparency of deep learning models and estimate their robustness accordingly. As an example use case, we apply our framework to facial expression recognition. We can show that the domain-specific evaluation is especially beneficial for challenging use cases such as facial expression recognition and provides application-grounded quality criteria that are not covered by standard evaluation methods. Our comparison of the domain-specific evaluation method with standard approaches thus shows that the quality of the expert knowledge is of great importance for assessing a model’s performance precisely.},
  keywords = {Application-Grounded Facial Expression Recognition}
}

@article{schwalbe2024_94996,
  title = {A comprehensive taxonomy for explainable artificial intelligence : a systematic survey of surveys on methods and concepts},
  author = {Schwalbe, Gesina and Finzel, Bettina},
  year = {2024},
  journal = {Data mining and knowledge discovery},
  volume = {38},
  number = {5},
  pages = {3043-3101},
  publisher = {Springer Science + Business Media B.V},
  address = {Dordrecht [u.a.]},
  issn = {1384-5810},
  doi = {10.1007/s10618-022-00867-8},
  url = {https://fis.uni-bamberg.de/handle/uniba/94996},
  language = {eng},
  abstract = {In the meantime, a wide variety of terminologies, motivations, approaches, and evaluation criteria have been developed within the research field of explainable artificial intelligence (XAI). With the amount of XAI methods vastly growing, a taxonomy of methods is needed by researchers as well as practitioners: To grasp the breadth of the topic, compare methods, and to select the right XAI method based on traits required by a specific use-case context. Many taxonomies for XAI methods of varying level of detail and depth can be found in the literature. While they often have a different focus, they also exhibit many points of overlap. This paper unifies these efforts and provides a complete taxonomy of XAI methods with respect to notions present in the current state of research. In a structured literature analysis and meta-study, we identified and reviewed more than 50 of the most cited and current surveys on XAI methods, metrics, and method traits. After summarizing them in a survey of surveys, we merge terminologies and concepts of the articles into a unified structured taxonomy. Single concepts therein are illustrated by more than 50 diverse selected example methods in total, which we categorize accordingly. The taxonomy may serve both beginners, researchers, and practitioners as a reference and wide-ranging overview of XAI method traits and aspects. Hence, it provides foundations for targeted, use-case-oriented, and context-sensitive future research.},
  keywords = {Explainable artificial intelligence, Interpretability, Taxonomy, Meta-analysis, Survey-of-surveys, Review}
}

@article{mohammed2024_93007,
  title = {Explaining and Evaluating Deep Tissue Classification by Visualizing Activations of Most Relevant Intermediate Layers},
  author = {Mohammed, Aliya and Geppert, Carol and Hartmann, Arnd and Kuritcyn, Petr and Bruns, Volker and Schmid, Ute and Wittenberg, Thomas and Benz, Michaela and Finzel, Bettina},
  year = {2024},
  pages = {229-232},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-930077},
  url = {https://fis.uni-bamberg.de/handle/uniba/93007},
  language = {eng},
  abstract = {Deep Learning-based tissue classification may support pathologists in analyzing digitized whole slide images. However, in such critical tasks, only approaches that can be validated by medical experts in advance to deployment, are suitable. We present an approach that contributes to making automated tissue classification more transparent. We step beyond broadly used visualizations for last layers of a convolutional neural network by identifying most relevant intermediate layers applying Grad-CAM. A visual evaluation by a pathologist shows that these layers assign relevance, where important morphological structures are present in case of correct class decisions. We introduce a tool that can be easily used by medical experts for such validation purposes for any convolutional neural network and any layer. Visual explanations for intermediate layers provide insights into a neural network’s decision for histopathological tissue classification. In future research also the context of the input data must be considered.},
  keywords = {Digital Pathology, Deep Learning, XAI, Evaluation, Grad-CAM, Intelligent user interface}
}

@article{holzinger2023_106517,
  title = {Toward human-level concept learning : Pattern benchmarking for AI algorithms},
  author = {Holzinger, Andreas and Saranti, Anna and Angerschmid, Alessa and Finzel, Bettina and Schmid, Ute and Mueller, Heimo},
  year = {2023},
  journal = {Patterns},
  volume = {4},
  number = {8},
  pages = {1-21},
  publisher = {Elsevier},
  address = {Amsterdam},
  issn = {2666-3899},
  doi = {10.1016/j.patter.2023.100788},
  url = {https://fis.uni-bamberg.de/handle/uniba/106517},
  language = {eng},
  abstract = {Artificial intelligence (AI) today is very successful at standard pattern-recognition tasks due to the availability of large amounts of data and advances in statistical data-driven machine learning. However, there is still a large gap between AI pattern recognition and human-level concept learning. Humans can learn amazingly well even under uncertainty from just a few examples and are capable of generalizing these concepts to solve new conceptual problems. The growing interest in explainable machine intelligence requires experimental environments and diagnostic/benchmark datasets to analyze existing approaches and drive progress in pattern analysis and machine intelligence. In this paper, we provide an overview of current AI solutions for benchmarking concept learning, reasoning, and generalization; discuss the state-of-the-art of existing diagnostic/ benchmark datasets (such as CLEVR, CLEVRER, CLOSURE, CURI, Bongard-LOGO, V-PROM, RAVEN, Kandinsky Patterns, CLEVR-Humans, CLEVRER-Humans, and their extension containing human language); and provide an outlook of some future research directions in this exciting research domain.}
}

@article{finzel2023_91640,
  title = {Generating Explanations for Conceptual Validation of Graph Neural Networks : An Investigation of Symbolic Predicates Learned on Relevance-Ranked Sub-Graphs},
  author = {Finzel, Bettina and Saranti, Anna and Angerschmid, Alessa and Tafler, David and Pfeifer, Bastian and Holzinger, Andreas},
  year = {2023},
  pages = {271–285},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-916401},
  url = {https://fis.uni-bamberg.de/handle/uniba/91640},
  language = {eng},
  abstract = {Graph Neural Networks (GNN) show good performance in relational data classification. However, their contribution to concept learning and the validation of their output from an application domain’s and user’s perspective have not been thoroughly studied. We argue that combining symbolic learning methods, such as Inductive Logic Programming (ILP), with statistical machine learning methods, especially GNNs, is an essential forward-looking step to perform powerful and validatable relational concept learning. In this contribution, we introduce a benchmark for the conceptual validation of GNN classification outputs. It consists of the symbolic representations of symmetric and non-symmetric figures that are taken from a well-known Kandinsky Pattern data set. We further provide a novel validation framework that can be used to generate comprehensible explanations with ILP on top of the relevance output of GNN explainers and human-expected relevance for concepts learned by GNNs. Our experiments conducted on our benchmark data set demonstrate that it is possible to extract symbolic concepts from the most relevant explanations that are representative of what a GNN has learned. Our findings open up a variety of avenues for future research on validatable explanations for GNNs.},
  keywords = {Graph neural networks (GNN), Explainable AI (xAI), Inductive logic programming (ILP), Symbolic AI, Kandinsky pattern (KP)}
}

@inproceedings{finzel2023_93116,
  title = {Domain-Specific Evaluation of Visual Explanations for Application-Grounded Facial Expression Recognition},
  author = {Finzel, Bettina and Rieger, Ines and Kuhn, Simon and Schmid, Ute Maria},
  year = {2023},
  booktitle = {Machine learning and knowledge extraction : 7th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2023, Benevento, Italy, August 29 – September 1, 2023, proceedings},
  pages = {31-44},
  publisher = {Springer Nature Switzerland},
  address = {Cham, Switzerland},
  isbn = {978-3-031-40836-6},
  doi = {10.1007/978-3-031-40837-3\_3},
  url = {https://fis.uni-bamberg.de/handle/uniba/93116},
  language = {eng},
  abstract = {Research in the field of explainable artificial intelligence has produced a vast amount of visual explanation methods for deep learning-based image classification in various domains of application. However, there is still a lack of domain-specific evaluation methods to assess an explanation’s quality and a classifier’s performance with respect to domain-specific requirements. In particular, evaluation methods could benefit from integrating human expertise into quality criteria and metrics. Such domain-specific evaluation methods can help to assess the robustness of deep learning models more precisely. In this paper, we present an approach for domain-specific evaluation of visual explanation methods in order to enhance the transparency of deep learning models and estimate their robustness accordingly. As an example use case, we apply our framework to facial expression recognition. We can show that the domain-specific evaluation is especially beneficial for challenging use cases such as facial expression recognition and provides application-grounded quality criteria that are not covered by standard evaluation methods. Our comparison of the domain-specific evaluation method with standard approaches thus shows that the quality of the expert knowledge is of great importance for assessing a model’s performance precisely.},
  keywords = {Application-Grounded Facial Expression Recognition}
}

@article{mohammed2022_93000,
  title = {Explaining and Evaluating Deep Tissue Classification by Visualizing Activations of Most Relevant Intermediate Layers},
  author = {Mohammed, Aliya and Geppert, Carol and Hartmann, Arnd and Kuritcyn, Petr and Bruns, Volker and Schmid, Ute and Wittenberg, Thomas and Benz, Michaela and Finzel, Bettina},
  year = {2022},
  journal = {Current directions in biomedical engineering},
  volume = {8},
  number = {2},
  pages = {229-232},
  publisher = {De Gruyter},
  address = {Berlin},
  issn = {2364-5504},
  doi = {10.1515/cdbme-2022-1059},
  url = {https://fis.uni-bamberg.de/handle/uniba/93000},
  language = {eng},
  abstract = {Deep Learning-based tissue classification may support pathologists in analyzing digitized whole slide images. However, in such critical tasks, only approaches that can be validated by medical experts in advance to deployment, are suitable. We present an approach that contributes to making automated tissue classification more transparent. We step beyond broadly used visualizations for last layers of a convolutional neural network by identifying most relevant intermediate layers applying Grad-CAM. A visual evaluation by a pathologist shows that these layers assign relevance, where important morphological structures are present in case of correct class decisions. We introduce a tool that can be easily used by medical experts for such validation purposes for any convolutional neural network and any layer. Visual explanations for intermediate layers provide insights into a neural network’s decision for histopathological tissue classification. In future research also the context of the input data must be considered.},
  keywords = {Digital Pathology, Deep Learning, XAI, Evaluation, Grad-CAM, Intelligent user interface}
}

@article{finzel2022_91634,
  title = {Generating Explanations for Conceptual Validation of Graph Neural Networks : An Investigation of Symbolic Predicates Learned on Relevance-Ranked Sub-Graphs},
  author = {Finzel, Bettina and Saranti, Anna and Angerschmid, Alessa and Tafler, David and Pfeifer, Bastian and Holzinger, Andreas},
  year = {2022},
  journal = {Künstliche Intelligenz : KI ; Forschung, Entwicklung, Erfahrungen ; Organ des Fachbereichs 1 Künstliche Intelligenz der Gesellschaft für Informatik e.V., GI},
  volume = {36},
  number = {3-4},
  pages = {271-285},
  publisher = {Springer},
  address = {Berlin},
  issn = {0933-1875},
  doi = {10.1007/s13218-022-00781-7},
  url = {https://fis.uni-bamberg.de/handle/uniba/91634},
  language = {eng},
  abstract = {Graph Neural Networks (GNN) show good performance in relational data classification. However, their contribution to concept learning and the validation of their output from an application domain’s and user’s perspective have not been thoroughly studied. We argue that combining symbolic learning methods, such as Inductive Logic Programming (ILP), with statistical machine learning methods, especially GNNs, is an essential forward-looking step to perform powerful and validatable relational concept learning. In this contribution, we introduce a benchmark for the conceptual validation of GNN classification outputs. It consists of the symbolic representations of symmetric and non-symmetric figures that are taken from a well-known Kandinsky Pattern data set. We further provide a novel validation framework that can be used to generate comprehensible explanations with ILP on top of the relevance output of GNN explainers and human-expected relevance for concepts learned by GNNs. Our experiments conducted on our benchmark data set demonstrate that it is possible to extract symbolic concepts from the most relevant explanations that are representative of what a GNN has learned. Our findings open up a variety of avenues for future research on validatable explanations for GNNs.},
  keywords = {Graph neural networks (GNN), Explainable AI (xAI), Inductive logic programming (ILP), Symbolic AI, Kandinsky pattern (KP)}
}

@inproceedings{rieger2022_58025,
  title = {CorrLoss : Integrating Co-Occurrence Domain Knowledge for Affect Recognition},
  author = {Rieger, Ines and Pahl, Jaspar and Finzel, Bettina and Schmid, Ute},
  year = {2022},
  booktitle = {2022 26th International Conference on Pattern Recognition (ICPR): Proceedings},
  pages = {798-804},
  publisher = {IEEE},
  isbn = {978-1-6654-9062-7},
  doi = {10.1109/ICPR56361.2022.9956319},
  url = {https://fis.uni-bamberg.de/handle/uniba/58025},
  language = {eng}
}

@article{schmid2022_55221,
  title = {Mutual Explanations for Cooperative Decision Making in Medicine},
  author = {Schmid, Ute and Finzel, Bettina},
  year = {2022},
  pages = {227-233},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-552219},
  url = {https://fis.uni-bamberg.de/handle/uniba/55221},
  language = {eng},
  abstract = {Exploiting mutual explanations for interactive learning is presented as part of an interdisciplinary research project on transparent machine learning for medical decision support. Focus of the project is to combine deep learning black box approaches with interpretable machine learning for classification of different types of medical images to combine the predictive accuracy of deep learning and the transparency and comprehensibility of interpretable models. Specifically, we present an extension of the Inductive Logic Programming system Aleph to allow for interactive learning. Medical experts can ask for verbal explanations. They can correct classification decisions and in addition can also correct the explanations. Thereby, expert knowledge can be taken into account in form of constraints for model adaption.},
  keywords = {Human-AI partnership, Inductive Logic Programming, Explanations as constraints}
}

@inproceedings{finzel2022_55380,
  title = {Deriving Temporal Prototypes from Saliency Map Clusters for the Analysis of Deep-Learning-based Facial Action Unit Classification},
  author = {Finzel, Bettina and Kollmann, Rene and Rieger, Ines and Pahl, Jaspar and Schmid, Ute},
  year = {2022},
  pages = {86-97},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-553809},
  url = {https://fis.uni-bamberg.de/handle/uniba/55380},
  language = {eng},
  abstract = {Reliably determining the emotional state of a person is a difficult task for both humans as well as machines. Automatic detection and evaluation of facial expressions is particularly important if people are unable to express their emotional state themselves, for example due to cognitive impairments. Identifying the presence of Action Units in a human’s face is a psychologically validated approach of quantifying which emotion is expressed. To automate the detection process of Action Units Neural Networks have been trained. However, the black-box nature of Deep Neural Networks provides no insight on the relevant features identified during the decision process. Approaches of Explainable Artificial Intelligence have to be applied to provide an explanation why the network came to a certain conclusion. In this work "Layer-Wise Relevance Propagation" (LRP) in combination with the meta analysis approach "Spectral Relevance Analysis" (SpRAy) is used to derive temporal prototypes from predictions in video sequences. Temporal prototypes provide an aggregated view on the prediction of the network by grouping together similar frames by considering relevance. Additionally, a specific visualization method for temporal prototypes is presented that highlights the most relevant areas for a prediction of an Action Unit. A quantitative evaluation of our approach shows that temporal prototypes aggregate temporal information well. The proposed method can be used to generate concise visual explanations for a sequence of interpretable saliency maps. Based on the above, this work shall provide the foundation for a new temporal analysis method as well as an explanation approach that is supposed to help researchers and experts to gain a deeper understanding of how the underlying network decides which Action Units are active in a particular emotional state.},
  keywords = {Temporal Prototypes, Facial Action Unit Detection, Layer-wise Relevance Propagation, Spectral Clustering, Affective Computing, Correlation Analysis}
}

@inproceedings{finzel2021_111166,
  title = {Explanation as a Process  : User-Centric Construction of Multi-level and Multi-modal Explanations},
  author = {Finzel, Bettina and Tafler, David E. and Scheele, Stephan M. and Schmid, Ute},
  year = {2021},
  booktitle = {KI 2021: Advances in Artificial Intelligence : 44th German Conference on AI, Virtual Event, September 27 – October 1, 2021, Proceedings},
  pages = {80–94},
  publisher = {Springer International Publishing},
  address = {Cham},
  series = {Lecture Notes in Computer Science},
  isbn = {978-3-030-87626-5},
  doi = {10.1007/978-3-030-87626-5\_7},
  url = {https://fis.uni-bamberg.de/handle/uniba/111166},
  language = {eng}
}

@inproceedings{deuschel2021_90482,
  title = {Uncovering the Bias in Facial Expressions},
  author = {Deuschel, Jessica and Finzel, Bettina and Rieger, Ines},
  year = {2021},
  booktitle = {Kolloquium Forschende Frauen 2020 - Gender in Gesellschaft 4.0 : Beiträge Bamberger Nachwuchswissenschaftlerinnen},
  pages = {15-42},
  publisher = {University of Bamberg Press},
  address = {Bamberg},
  isbn = {978-3-86309-853-7},
  doi = {10.20378/irb-90482},
  url = {https://fis.uni-bamberg.de/handle/uniba/90482},
  language = {eng},
  abstract = {Over the past decades the machine and deep learning community has celebrated great achievements in challenging tasks such as image classification. The deep architecture of artificial neural networks together with the plenitude of available data makes it possible to describe highly complex relations. Yet, it is still impossible to fully capture what the deep learning model has learned and to verify that it operates fairly and without creating bias, especially in critical tasks, for instance those arising in the medical field. One example for such a task is the detection of distinct facial expressions, called Action Units, in facial images.
Considering this specific task, our research aims to provide transparency regarding bias, specifically in relation to gender and skin color. We train a neural network for Action Unit classification and analyze its performance quantitatively based on its accuracy and qualitatively based on heatmaps. A structured review of our results indicates that we are able to detect bias. Even though we cannot conclude from our results that lower classification performance emerged solely from gender and skin color bias, these biases must be addressed, which is why we end by giving suggestions on how the detected bias can be avoided.},
  keywords = {-}
}

@inproceedings{finzel2021_55037,
  title = {Deriving Temporal Prototypes from Saliency Map Clusters for the Analysis of Deep-Learning-based Facial Action Unit Classification},
  author = {Finzel, Bettina and Kollmann, Rene and Rieger, Ines and Pahl, Jaspar and Schmid, Ute},
  year = {2021},
  booktitle = {Proceedings of the LWDA 2021 Workshops: FGWM, KDML, FGWI-BIA, and FGIR},
  pages = {86-97},
  publisher = {RWTH Aachen},
  address = {Aachen, Germany},
  url = {https://fis.uni-bamberg.de/handle/uniba/55037},
  language = {eng},
  abstract = {Reliably determining the emotional state of a person is a difficult task for both humans as well as machines. Automatic detection and evaluation of facial expressions is particularly important if people are unable to express their emotional state themselves, for example due to cognitive impairments. Identifying the presence of Action Units in a human’s face is a psychologically validated approach of quantifying which emotion is expressed. To automate the detection process of Action Units Neural Networks have been trained. However, the black-box nature of Deep Neural Networks provides no insight on the relevant features identified during the decision process. Approaches of Explainable Artificial Intelligence have to be applied to provide an explanation why the network came to a certain conclusion. In this work "Layer-Wise Relevance Propagation" (LRP) in combination with the meta analysis approach "Spectral Relevance Analysis" (SpRAy) is used to derive temporal prototypes from predictions in video sequences. Temporal prototypes provide an aggregated view on the prediction of the network by grouping together similar frames by considering relevance. Additionally, a specific visualization method for temporal prototypes is presented that highlights the most relevant areas for a prediction of an Action Unit. A quantitative evaluation of our approach shows that temporal prototypes aggregate temporal information well. The proposed method can be used to generate concise visual explanations for a sequence of interpretable saliency maps. Based on the above, this work shall provide the foundation for a new temporal analysis method as well as an explanation approach that is supposed to help researchers and experts to gain a deeper understanding of how the underlying network decides which Action Units are active in a particular emotional state.},
  keywords = {Temporal Prototypes, Facial Action Unit Detection, Layer-wise Relevance Propagation, Spectral Clustering, Affective Computing, Correlation Analysis}
}

@misc{achter2021_50041,
  title = {Uni.vers Forschung : das Magazin der Otto-Friedrich-Universität Bamberg},
  author = {Achter, Patricia and Braches-Chyrek, Rita and Düchs, Martin and Färber, Tanja and Finzel, Bettina and Ganesch, Franziska and Graichen, Joanna and Horn, Matthias and Oehler, Andreas and Posegga, Oliver and Proaño Acosta, Christian and Scheltjens, Werner and Schmid, Ute and Schneider, Pauline and Staake, Thorsten and Steins-Löber, Sabine and Stingl, Carlo and Stöckl, Andreas and Struck, Olaf and Velde, Caroline van der and Wolstein, Jörg},
  editor = {Fischbach, Kai},
  year = {2021},
  number = {April 2021, Mensch und Maschine - Bamberger Beiträge zur Digitalisierung = Man and Machine - Bamberg Scholars on Digitalisation},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  issn = {1618-9019},
  urn = {urn:nbn:de:bvb:473-irb-500418},
  url = {https://fis.uni-bamberg.de/handle/uniba/50041},
  language = {deu},
  abstract = {Wenn Menschen die Technik weiterentwickeln, entstehen ständig neue Möglichkeiten und Herausforderungen. Die Gesellschaft ist gefragt mitzudiskutieren, wie eine immer weiter digitalisierte Welt künftig aussehen könnte. Und die Forschung hilft dabei, dass sich jede und jeder Einzelne eine fundierte Meinung bilden kann. In dieser Ausgabe schreiben die Autorinnen und Autoren über den Einsatz der Technik im Alltag, die Zusammenarbeit von Menschen und Maschinen, das Pro und Contra der Digitalisierung.},
  keywords = {Digitalisierung, Technik, KI, Gesellschaft, App, Internet, Arbeitsmarkt, Geschichte, Psychologie, Architektur, Kinder, Medien, Finanzen, Gesundheit}
}

@incollection{deuschel2021_50304,
  title = {Uncovering the Bias in Facial Expressions},
  author = {Deuschel, Jessica and Finzel, Bettina and Rieger, Ines},
  year = {2021},
  pages = {1-28},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  doi = {10.20378/irb-50304},
  urn = {urn:nbn:de:bvb:473-irb-503044},
  url = {https://fis.uni-bamberg.de/handle/uniba/50304},
  language = {eng},
  abstract = {Over the past decades the machine and deep learning community has celebrated great achievements in challenging tasks such as image classification. The deep architecture of artificial neural networks together with the plenitude of available data makes it possible to describe highly complex relations. Yet, it is still impossible to fully capture what the deep learning model has learned and to verify that it operates fairly and without creating bias, especially in critical tasks, for instance those arising in the medical field. One example for such a task is the detection of distinct facial expressions, called Action Units, in facial images. Considering this specific task, our research aims to provide transparency regarding bias, specifically in relation to gender and skin color. We train a neural network for Action Unit classification and analyze its performance quantitatively based on its accuracy and qualitatively based on heatmaps. A structured review of our results indicates that we are able to detect bias. Even though we cannot conclude from our results that lower classification performance emerged solely from gender and skin color bias, these biases must be addressed, which is why we end by giving suggestions on how the detected bias can be avoided.},
  keywords = {Facial Expressions, Bias, Neural Networks, Deep Learning, Facial Action Units}
}

@misc{schwalbe2021_51285,
  title = {XAI Method Properties : A (Meta-)study},
  author = {Schwalbe, Gesina and Finzel, Bettina},
  year = {2021},
  number = {Submitted on 15 May 2021},
  url = {https://fis.uni-bamberg.de/handle/uniba/51285},
  language = {eng}
}

@inproceedings{rauscher2021_50050,
  title = {Kolloquium Forschende Frauen 2020 : Gender in Gesellschaft 4.0 : Beiträge Bamberger Nachwuchswissenschaftlerinnen},
  author = {Rauscher, Judith and Deuschel, Jessica and Finzel, Bettina and Rieger, Ines and Wagner, Anne and Velde, Caroline van der and Oehlhorn, Caroline and Hlukhovych, Adrianna},
  editor = {Rauscher, Judith and Hess, Mona and Schütz, Astrid and Szczepaniak, Renata},
  year = {2021},
  publisher = {University of Bamberg Press},
  address = {Bamberg},
  series = {Forschende Frauen in Bamberg},
  isbn = {978-3-86309-853-7},
  doi = {10.20378/irb-50050},
  urn = {urn:nbn:de:bvb:473-irb-500505},
  url = {https://fis.uni-bamberg.de/handle/uniba/50050},
  language = {other},
  abstract = {Die Buchreihe „Forschende Frauen in Bamberg“ begleitet das gleichnamige Forschungskolloquium der Frauenbeauftragten der Otto-Friedrich-Universität Bamberg.
Das Thema für 2020 ist "Gender in Gesellschaft 4.0".},
  keywords = {Bias, Facial Expressions, Facial Action Units, Deep Learning, Neural Network, Skin color, Gender, Layer-wise Relevance Propagation, Machine Learning, Artificial Intelligence, FairAI, Interpretable AI, IT-Fachkräfte, Arbeitgeberattraktivität, Kano-Analyse, Gender, Medien, Geschlechtsidentitäten, Cyberfeminismus, Neuer Materialismus, kritischer Posthumanismus, I-GENDO, obesity, mHealth, gender, psychological therapy, I-GENDO, Adipositas, mHealth, gender, Psychotherapie, Digitalisierung, Berufliche Bildung, Berufsschule, Geschlechterunterschiede, Digitale Medien, Unterricht, Tablet}
}

@article{bruckert2020_57747,
  title = {The Next Generation of Medical Decision Support : A Roadmap Toward Transparent Expert Companions},
  author = {Bruckert, Sebastian and Finzel, Bettina and Schmid, Ute},
  year = {2020},
  journal = {Frontiers in artificial intelligence},
  volume = {3},
  publisher = {Frontiers Media},
  address = {Lausanne},
  issn = {2624-8212},
  doi = {10.3389/frai.2020.507973},
  url = {https://fis.uni-bamberg.de/handle/uniba/57747},
  language = {eng},
  abstract = {Increasing quality and performance of artificial intelligence (AI) in general and machine learning (ML) in particular is followed by a wider use of these approaches in everyday life. As part of this development, ML classifiers have also gained more importance for diagnosing diseases within biomedical engineering and medical sciences. However, many of those ubiquitous high-performing ML algorithms reveal a black-box-nature, leading to opaque and incomprehensible systems that complicate human interpretations of single predictions or the whole prediction process. This puts up a serious challenge on human decision makers to develop trust, which is much needed in life-changing decision tasks. This paper is designed to answer the question how expert companion systems for decision support can be designed to be interpretable and therefore transparent and comprehensible for humans. On the other hand, an approach for interactive ML as well as human-in-the-loop-learning is demonstrated in order to integrate human expert knowledge into ML models so that humans and machines act as companions within a critical decision task. We especially address the problem of Semantic Alignment between ML classifiers and its human users as a prerequisite for semantically relevant and useful explanations as well as interactions. Our roadmap paper presents and discusses an interdisciplinary yet integrated Comprehensible Artificial Intelligence (cAI)-transition-framework with regard to the task of medical diagnosis. We explain and integrate relevant concepts and research areas to provide the reader with a hands-on-cookbook for achieving the transition from opaque black-box models to interactive, transparent, comprehensible and trustworthy systems. To make our approach tangible, we present suitable state of the art methods with regard to the medical domain and include a realization concept of our framework. The emphasis is on the concept of Mutual Explanations (ME) that we introduce as a dialog-based, incremental process in order to provide human ML users with trust, but also with stronger participation within the learning process.},
  keywords = {explainable artificial intelligence, interactive ML, interpretability, trust, medical diagnosis, medical decision support, companion}
}

@inproceedings{rieger2020_49224,
  title = {Verifying Deep Learning-based Decisions for Facial Expression Recognition},
  author = {Rieger, Ines and Kollmann, Rene and Finzel, Bettina and Seuß, Dominik and Schmid, Ute},
  year = {2020},
  booktitle = {ESANN 2020 proceedings},
  pages = {139-144},
  isbn = {978-2-87587-074-2},
  url = {https://fis.uni-bamberg.de/handle/uniba/49224},
  language = {eng}
}

@article{finzel2020_49223,
  title = {Korrigierbares maschinelles Lernen mithilfe wechselseitiger Erklärungen am Beispiel der Medizin},
  author = {Finzel, Bettina},
  year = {2020},
  journal = {Frauen machen Informatik},
  number = {44},
  pages = {14-17},
  address = {Bonn},
  issn = {2364-6152},
  url = {https://fis.uni-bamberg.de/handle/uniba/49223},
  language = {deu}
}

@article{bruckert2020_49222,
  title = {The Next Generation of Medical Decision Support : A Roadmap Toward Transparent Expert Companions},
  author = {Bruckert, Sebastian and Finzel, Bettina and Schmid, Ute},
  year = {2020},
  number = {507973},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  url = {https://fis.uni-bamberg.de/handle/uniba/49222},
  language = {eng},
  abstract = {Increasing quality and performance of artificial intelligence (AI) in general and machine learning (ML) in particular is followed by a wider use of these approaches in everyday life. As part of this development, ML classifiers have also gained more importance for diagnosing diseases within biomedical engineering and medical sciences. However, many of those ubiquitous high-performing ML algorithms reveal a black-box-nature, leading to opaque and incomprehensible systems that complicate human interpretations of single predictions or the whole prediction process. This puts up a serious challenge on human decision makers to develop trust, which is much needed in life-changing decision tasks. This paper is designed to answer the question how expert companion systems for decision support can be designed to be interpretable and therefore transparent and comprehensible for humans. On the other hand, an approach for interactive ML as well as human-in-the-loop-learning is demonstrated in order to integrate human expert knowledge into ML models so that humans and machines act as companions within a critical decision task. We especially address the problem of Semantic Alignment between ML classifiers and its human users as a prerequisite for semantically relevant and useful explanations as well as interactions. Our roadmap paper presents and discusses an interdisciplinary yet integrated Comprehensible Artificial Intelligence (cAI)-transition-framework with regard to the task of medical diagnosis. We explain and integrate relevant concepts and research areas to provide the reader with a hands-on-cookbook for achieving the transition from opaque black-box models to interactive, transparent, comprehensible and trustworthy systems. To make our approach tangible, we present suitable state of the art methods with regard to the medical domain and include a realization concept of our framework. The emphasis is on the concept of Mutual Explanations (ME) that we introduce as a dialog-based, incremental process in order to provide human ML users with trust, but also with stronger participation within the learning process.},
  keywords = {explainable artificial intelligence, interactive ML, interpretability, trust, medical diagnosis, medical decision support, companion}
}

@article{schmid2020_47120,
  title = {Mutual Explanations for Cooperative Decision Making in Medicine},
  author = {Schmid, Ute and Finzel, Bettina},
  year = {2020},
  journal = {Künstliche Intelligenz : KI ; Forschung, Entwicklung, Erfahrungen},
  volume = {34},
  number = {2},
  pages = {227-233},
  publisher = {Springer},
  address = {Berlin ; Heidelberg},
  issn = {1610-1987},
  doi = {10.1007/s13218-020-00633-2},
  url = {https://fis.uni-bamberg.de/handle/uniba/47120},
  language = {eng},
  abstract = {Exploiting mutual explanations for interactive learning is presented as part of an interdisciplinary research project on transparent machine learning for medical decision support. Focus of the project is to combine deep learning black box approaches with interpretable machine learning for classification of different types of medical images to combine the predictive accuracy of deep learning and the transparency and comprehensibility of interpretable models. Specifically, we present an extension of the Inductive Logic Programming system Aleph to allow for interactive learning. Medical experts can ask for verbal explanations. They can correct classification decisions and in addition can also correct the explanations. Thereby, expert knowledge can be taken into account in form of constraints for model adaption.},
  keywords = {Human-AI partnership, Inductive Logic Programming, Explanations as constraints}
}

@techreport{fortsch2017_42406,
  title = {Unternehmenskulturen im Wandel - Karriereentwicklung im Informatikbereich. Ergebnisse des Forschungsprojektes "Alumnae Tracking"},
  author = {Förtsch, Silvia and Gärtig-Daugs, Anja and Schmid, Ute},
  year = {2017},
  publisher = {opus},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-opus4-499108},
  url = {https://fis.uni-bamberg.de/handle/uniba/42406},
  language = {deu},
  abstract = {Unternehmenskulturen im Wandel beinhaltet quantitative Ergebnisse aus der Studierenden- und Ehemaligenbefragung sowie qualitative Ergebnisse aus der Unternehmensbefragung im Forschungsprojekt "Alumnae Tracking".}
}

